ABSTRACT CD8 T cells are important contributors to the pathology observed in many autoimmune diseases, including type 1 diabetes. Despite the importance of autoreactive CD8 T cells, much fundamental knowledge is lacking in several major areas, including i) the properties of class I MHC alleles that confer susceptibility or resistance to disease; ii) the determinants that underlie the promiscuous antigen recognition often displayed by autoreactive T cells; and iii) the contribution of the CD8?? co-receptor in facilitating autoimmune processes. These are the critical knowledge gaps that we seek to fill, and we will do so using a multidisciplinary approach and a collection of innovative strategies, reagents, and animal models. In Aim 1, we will define the biochemical and structural features and in vivo peptide presentation properties of human class I MHC alleles associated with susceptibility or resistance to type 1 diabetes. These studies will focus on HLA-B*39:06, B*39:01, and B*38:01, a closely related group of class I MHC alleles that are differentially associated with disease. We will test the hypothesis that the differential ability of these three alleles to mediate disease is the consequence of differential binding and presentation of certain disease-relevant beta cell peptides. Our newly developed NOD mouse models expressing human insulin or human IGRP will be central to this aim. In Aim 2, we will use the T cell receptors from a set of promiscuous and autoreactive CD8 T cell clones in structural and mutagenesis studies to define the mechanisms that underlie T cell receptor cross-reactivity. Among other approaches, we will take advantage of novel high-throughput assays that we have developed both to identify T cell epitopes and map protein binding interfaces. In Aim 3, we will use biochemical and biophysical approaches and in vivo models to better understand the interaction between CD8?? and peptide-MHC. Despite the requirement of CD8?? for CD8 T cell selection and peripheral activation, and its potential as a therapeutic target, crystallographic analysis of its interaction with pMHC is limited to a single structure of murine CD8?? with H-2Dd. Our own findings from mouse models we have recently characterized suggest that human CD8?? may be interacting with class I MHC in a manner not revealed by the single available structure. While autoreactive CD8 T cells under study here are likely to exhibit some unique features due to their necessity to escape tolerance mechanisms, much of what will be learned will be fundamental knowledge that is generally applicable to CD8 T cell biology and that will help guide the future development of more effective approaches to manipulate and harness the immune system in health and disease.